Semantic Web Service Discovery in the OWLS IDE

1
Semantic Web Service Discovery in the OWL-S IDE

• ???
• 2007/06/13

2
Outline

• Introduction
• Problem
• Solution
• OWL-S
• OWL-S IDE
• Extending UDDI using OWL-S
• Extending UDDI API
• Matching Algorithm
• Conclusion

3
Introduction

• The increasing availability of web services
  necessitates efficient discovery and execution
  framework.
• It is becoming increasingly difficult for service
  requester to automatically find service providers
  that satisfy its requirements.

4
Problem

• Some of these difficulties are attributed to the
  use of XML.
• Two syntactically identical XML descriptions may
  have very different meaning
• Two syntactically different XML descriptions may
  have the same meaning

5
Problem (cont.)

• UDDI is an existing industry standard developed
  to solve the web service discovery problem.
• First limitation is its search mechanism.
• Second shortcoming is the usage of XML to
  describe its data model

6
Solution

• In addition to the UDDI like classification-based
  search mechanism, OWL-S also provides
  capability-based search mechanism
• Using capability-based search we can discovery
  web services based on
• the inputs and preconditions that need to be
  satisfied
• outputs and effects that need to be produced.

7
OWL-S

• a Web Services ontology that specifies a
  conceptual framework for describing semantic web
  services
• a language that enriches Web Services
  descriptions with semantic information

8
OWL-S

• OWL-S is characterized by three modules
• Profile describes capabilities of Web Services
  as well as additional features (e.g. inputs,
  outputs, preconditions and effects) of web
  services
• Process Model provides a description of the
  activity of the Web Service provider from which
  the Web Service requester can derive the
  interaction
• Grounding a description of how abstract
  information exchanges described in the Process
  Model are mapped onto actual messages that the
  provider and the requester exchange.

9
OWL-S IDE
10
OWL-S IDE (cont.)

• Two difference approaches
• code-driven approach the web service is
  implemented using Java or any other programming
  language, and the OWL-S descriptions are derived
  from the code using WSDL2OWL-S converter
• Good legacy systems can easily expose them as
  web services
• Poor missing semantic and control flow
  information

11
OWL-S IDE (cont.)

• Two difference approaches
• model-driven approach starts with the OWL-S
  specification of the functionalities that the
  developer expects from the Web service, and the
  specification of the Web service interaction
  process, and ends with a partial generation of
  the (Java) code of the Web service.

12
Extending UDDI using OWL-S

• In order to take advantage of the semantic
  matching and the capability-based search provided
  by OWL-S we need to embed OWL-S Profile
  information inside UDDI advertisements.

13
Extending UDDI using OWL-S

• Mapping between OWL-S Profile and UDDI
• one-to-one mapping if an OWL-S profile element
  has a corresponding UDDI element
• T-Model based mapping OWL-S profile elements
  with no corresponding UDDI elements

14
Extending UDDI using OWL-S
15
Extending UDDI API

• The UDDI registry is extended with capability
  port to receive and process semantic queries.

16
Matching Algorithm (cont.)

• When a request is submitted
• First matching the outputs of the request against
  the outputs of the published advertisement
• Then the inputs of the request are matched
  against the inputs of the advertisements matched

17
Matching Algorithm (cont.)

• subsumption mechanism a flexible matching
  mechanism based on the OWL

18
Conclusion

• the challenges posed by existing web service
  standards to automatically discover and interact
• Provide algorithms for the efficient use of OWL-S
  ontologies in UDDI

19
(No Transcript)